导　　师： 屈卫东

学科专业： 081101

授予学位： 硕士

作　　者： ;

机构地区： 上海交通大学

摘　　要： 燃油调节系统是航空发动机控制系统的重要组成部分。随着航空发动机控制精度要求的提高,燃油调节系统的性能优劣就显得尤为重要。此外,在实际中经常会遇到的性能故障,如超温、超转、转速摆动等,往往都是由于调节器故障引起的。因此,对航空发动机燃调系统进行故障诊断方面的研究具有非常重要的工程意义。 针对航空发动机燃调系统工作机制错综复杂的特点,如果不对发动机进行彻底理解和学习就无法实现相应故障的检测与诊断。本文提出基于符号有向图SDG/(Signed Directed Graph/)的知识模型,该模型能够表达复杂的因果关系,具有包容大规模潜在信息的能力,能够针对航空发动机结构复杂、故障原因多的特点具体的解决问题。相对于其他的故障诊断算法,SDG简单直观,方便应用于工程实际。本文从工程角度出发,根据工程经验,针对于某型航空发动机的燃油控制系统建立了对应的SDG知识系统模型,定制了相应的故障推理算法,并验证了算法的有效性。 在对燃调系统进行整体故障分析的基础之上,本文对其关键部件电液伺服阀进行了故障诊断方面的研究。电液伺服阀作为燃调系统的关键部件,其性能好坏直接影响到系统的稳定性和精度。本文具体分析了电液伺服阀的工作原理,建立了液压仿真模型,并通过调整模型参数进行故障模拟,进而获取故障数据;分别采用支持向量机/(support vector machine,SVM/)、主元分析法-支持向量机/(principal component analysis-support vector machine,PCA-SVM /)和小波包能量特征向量-支持向量机/(wavelet packet energy eigenvector-support vector machine ,WPEE-SVM/)/)进行故障分类。经综合比较分类精度和速度,选择PCA-SVM作为最佳诊断算法,同时也证明了PCA-SVM算法的有效性。 最后本文介绍了发动机一体化仿真平台。对于该仿真平台的结构和功能进行了一定程度的简要描述。 Fuel control system is quite an important component of a whole aero-engine control system. As the accuracy and stability of an aero-engine control system is increased these years, the performance of fuel-control system is quite important to help achieve that point. Furthermore, many common problems are caused by fuel-control system. So research on fault diagnosis of fuel control system has import engineering significance. The working mechanism of aero-engine is complicated, so we can not figure out failure without a deep understanding of it. For this special situation mentioned above, knowledge model based on SDG/(signed Directed Graph/) can express complex relationship. SDG has a huge potential in covering large-scale information and can solve well for the problem. Comparing with other method, SDG is simple and audio-visual, which make it quite easy for practical use. Based on the engineering experiences, we establish a SDG expert system model for fuel control system. Then we work out its realization and verify the effectiveness of it. As a core component of fuel-control system, the working performance of electro-hydraulic servo valve directly affects the stability and accuracy of the whole servo system. Research on fault diagnosis of electro-hydraulic servo valve has important engineering significance. Besides the fault diagnosis of the whole fuel control system, this paper also focuses on the fault analysis of electro-hydraulic servo valve. For a certain twin flapper-nozzle electro-hydraulic servo valve, analysis its work mechanism, build its hydraulic model and alter some parameters of it to simulate fault. On this basis, employ SVM,PCA-SVM and WPEE-SVM to train and classify the fault data. After comparison, verify the effectiveness of algorithm and choose PCA-SVM as the best method. At last, we introduce the simulation platform of aero-engine. We also explain the structure and function of the platform in brief.

关 键 词： 航空发动机 符号有向图 故障诊断 燃调系统 电液伺服阀 主元分析法 小波包特征向量 支持向量机

分 类 号： [V263.6]

领　　域： [航空宇航科学与技术]